© 2018 Kurt Pfeifer.

Rationale for Hybrid CAD Algorithm

The hybrid algorithm included in this guide is an amalgamation of the ACC/AHA [1] ESC/ESA [2], and CCS [3] guidelines and other high-quality studies. The ACC/AHA and ESC/ESA guidelines were published in 2014, prior to several practice-changing studies in perioperative medicine. The CCS guidelines were published in 2017 and included some of these studies, but since early 2017, several other important perioperative studies have been completed. This algorithm is the author’s suggestion for combining all the guidelines and most recent literature. This algorithm has not been prospectively validated. A step-by-step explanation follows:

  1. Emergency surgery: All guidelines and experts agree that emergent surgery should not be delayed for further cardiac risk stratification. The ACC/AHA define emergent surgery as being required within 6 hours to prevent loss of life or limb.

  2. Symptoms of ischemic disease: The ACC/AHA, ESC/ESA, and CCS algorithms include assessment for acute coronary syndrome or unstable/active cardiac conditions as an early step. However, none of the available guideline algorithms contain an assessment for non-acute cardiovascular symptoms. While it may not always trigger further testing or management, a patient who reports exertional chest pain or other symptoms that the clinician considers suggestive of myocardial ischemia would warrant further evaluation of these symptoms in the general setting.[4] If indicated in the general setting, it should certainly be considered prior to an elevated risk scenario such as major surgery. Of course, the timeline for surgery and overall clinical context must be factored in, but this symptomatology must be prominently factored into perioperative planning.

  3. RCRI of 0 early in algorithm for rapid rule-out of need for further evaluation: Asymptomatic patients with an RCRI of 0 have consistently been shown to have low risk of major adverse cardiovascular events (MACE). The RCRI is quick and easy as well. Therefore, it is clinically efficient to apply it early in the assessment for ischemic risk because if it is 0 (and the patient has no symptoms), one can safely conclude that further coronary disease evaluation would not benefit the patient. Although it was not designed/validated for low-risk/outpatient surgery, a score of 0 is valid for low-risk stratification in any surgical patient. The ACC/AHA defines an RCRI of 0 or 1 as low risk, and the ESC/ESA actually recommends no further evaluation for an RCRI of 2 or less. In modern cohorts, even an RCRI of 1 carries a MACE risk of well over 1%, which is the level the ACC/AHA defines as “elevated risk.” [3,5] Thus, including an RCRI of 1 as low-risk is incongruent with the remainder of the ACC/AHA algorithm. This is supported by the CCS guideline which uses only an RCRI of 0 plus age <65 years as criterion defining low-risk requiring no further assessment. The addition of age <65 years is based upon data from the VISION trial showing increased risk of MACE in patients older than 65 years, even if they had no other clinical risk factors. [6]

  4. Coronary event or intervention in past year: This correlates with the ACC/AHA and ESC/ESA inclusion of acute coronary syndromes (ACS) as a stop point within their algorithms. The author expanded this to include any coronary intervention because such patients require additional consideration of the timing of a surgical procedure relative to when they underwent a coronary intervention. Furthermore, ACC appropriate use criteria for stable CAD provide specific recommendations for when repeat coronary evaluation (with stress testing, coronary CT, or coronary angiography) is indicated in a patient with previous coronary evaluation or intervention. These state that repeat testing is not indicated within a year of previous evaluation/intervention in a clinically stable patient. [4] Combining ACS and recent intervention early in the algorithm, allow for these patients to be more efficiently risk stratified.

  5. Normal coronary evaluation in the past year: As described above, ACC appropriate use criteria state that repeat testing is not indicated within a year of previous evaluation/intervention in a clinically stable patient. [4] This step is included before more complicated surgical risk assessments because it may allow for more efficient completion of cardiac ischemic disease risk stratification.

  6. Estimated risk of MACE: This step correlates with the ACC/AHA algorithm, which advises use of the RCRI, NSQIP MICA index or ACS surgical risk calculator. Neither the ESC/ESA or CCS include the use of the NSQIP MICA index or the ACS surgical risk calculator as options for MACE estimation; instead, they utilize the RCRI because they point out the lack of external validations of these tools. The hybrid algorithm includes only the NSQIP MICA index and ACS surgical risk calculator at this step because as described above, the ability of the RCRI to discriminate levels of risk beyond an RCRI of 0 is quite limited. [3,5] For instance, the CCS guideline’s pooled risk estimates from recent external validations of the RCRI show respective MACE risks of: 3.9% for RCRI 0, 6.0% for RCRI 1, 10.1% for RCRI 2, and 15% for RCRI 3 or higher. Delineating risk between these large risk percentage intervals is extremely challenging at best. (Note: While a MACE rate of 3.9% for an RCRI of 0 might challenge the idea of using an RCRI of 0 as a criterion of low-risk, most of these studies employed routine troponin monitoring and thus had higher reporting of MI. Within this context, this rate of MACE appears to be consistent with baseline level of risk. [3]) As described above, an RCRI of 0 can be sufficient to “escape the algorithm” without exerting much more effort in risk stratification. However, if it isn’t 0, the NSQIP MICA index and ACS surgical risk calculators likely provide a more specific assessment of the patient’s risk since they combine both surgical and patient risk factors. Indeed, C-stats suggest these two have superior predictive capacity. [7,8] While neither is perfect (as illustrated by smaller, single-institution studies), a recent study confirmed that they likely provide better risk stratification than the RCRI. [9] In the end, the purpose of these tools is not to perfectly predict risk but rather stratify/categorize a patient’s risk for informed decision-making and perioperative care planning. Toward that end, these tools likely provide a more refined assessment.

  7. Functional capacity assessment using DASI: Functional capacity assessment is included in both the ACC/AHA and ESC/ESA algorithms, in which determination that the patient can achieve >4 METs of exertion leads to the recommendation for no further testing. The CCS guideline specifically leaves out functional capacity assessment, citing both the paucity and variable findings of studies evaluating its utility in perioperative risk prediction. Since the publication of the CCS guideline, the METS trial was completed and found that of subjective functional capacity assessment, functional capacity assessment with DASI, cardiopulmonary exercise testing, and BNPs, only functional capacity assessment was predictive of the primary outcome of 30-day death+MI. [10] Subsequent analyses of the METS trial are forthcoming and will likely provide recommendations for a specific DASI score cut-off predictive of low-risk. Until those are available, it is recommended to use the previously validated METs estimation produced from the DASI score and available through the calculator in the hyperlink.

  8. Surgery time-sensitive: Although the ACC/AHA guideline provides definitions for urgent (required within 24 hours) and time-sensitive (required within 6 weeks) surgery, they do not provide specific recommendations or algorithm components for these categories. The ESC/ESA algorithm has “urgent surgery” (rather than emergency surgery in the ACC/AHA guideline) as its first step and recommends no further testing in this scenario. The CCS guideline separates emergency and urgent surgery, recommending no further evaluation of any kind for emergency surgery and for urgent surgery, no further evaluation except in the case of unstable cardiac conditions or suspected severe obstructive cardiac disease (eg, aortic stenosis) or pulmonary hypertension. Given the unlikelihood of accomplishment of CAD interventions (ie, stenting or bypass surgery with their necessary periods of uninterrupted dual anti-platelet therapy) within 6 weeks, further coronary evaluation is not recommended for time-sensitive surgeries in asymptomatic patients.

  9. Coronary evaluation: All 3 major guidelines acknowledge that available literature has not demonstrated benefit from preoperative, prophylactic (in asymptomatic patients) coronary revascularization. The ACC/AHA and ESC/ESA still suggest that noninvasive coronary evaluation can be considered if it will change management. Other than the dismissed option of coronary intervention, the primary way it could change management would be to provide additional information for shared decision-making by the patient and surgeon. If concerned about risk, and the options of no surgery or a lower-risk procedure are viable, noninvasive evaluation could potentially help patients and surgeons in making the decision. However, clinicians must remember that noninvasive coronary evaluation is not perfect: positive predictive value <40% with a negative predictive value of >90%. The latter is also much lower in patients with high pre-test probability for CAD. For these reasons, the CCS does not suggest preoperative noninvasive coronary evaluation in any situation. This hybrid algorithm has stress-testing as a consideration but should only be ordered after full consideration of the above limitations.

  10. BNP for preoperative risk stratification: The ACC/AHA states only that the role of preoperative cardiac biomarkers remains unclear. The ESC/ESA guideline acknowledged the uncertainty of how to utilize the results of preoperative cardiac biomarkers but does say that preoperative BNPs may be considered in high-risk patients: METs ≤4 or RCRI >1 for vascular surgery or >2 for nonvascular surgery. The CCS guideline goes much further and recommends obtaining preoperative BNP in any patient with: age ≥65 years, RCRI ≥1, or age ≥45 years + known cardiovascular disease. The CCS cites strong evidence showing that BNPs provide significant improvement in risk estimation when combined with clinical assessment. Most of these studies have utilized RCRI or very similar criteria for clinical assessment. For those with BNP ≥92 or NT-proBNP ≥300, the CCS guideline recommends daily ECG & troponin for 48-72 hours after surgery. In this approach, the CCS uses the BNPs to identify higher risk patients but then just accepts the risk and uses heightened postoperative monitoring to detect myocardial ischemia. This site’s hybrid algorithm incorporates BNPs in the same step as stress testing because BNPs and stress testing have comparable negative and positive predictive values. In both cases, a normal value is consistent with a level of risk unlikely to be altered by further evaluation, while an abnormal test suggests the patient may be at increased risk for postoperative CV events. In such cases, there is still no evidence that revascularization helps, but since the test was ordered with the understanding that the results would alter perioperative management, the hybrid algorithm suggests referral to cardiology for additional input for perioperative optimization. Prior to utilizing BNP for such a purpose, one should determine if all members of the perioperative team (anesthesiologist, surgeon, and cardiologist) are agreeable to this approach. In other words, the individual performing the preop evaluation shouldn’t rely on a normal BNP to rule-out the need for further testing if the anesthesiologist or surgeon will not be comfortable with this and would want noninvasive coronary evaluation instead.

References:

  1. Fleisher LA et al. J Am Coll Cardiol. 2014;64(22):77-137.

  2. Kristensen SD et al. Eur Heart J. 2014;35(35):2383-431.

  3. Duceppe E et al. Can J Cardiol. 2017 Jan;33(1):17-32.

  4. Wolk MJ et al. J Am Coll Cardiol. 2014;63(4):380-406.

  5. Davis C et al. Can J Anaesth. 2013;60(9):855-63.

  6. Devereaux PJ et al. JAMA. 2012;307(21):2295-304.

  7. Gupta PK et al. Circulation. 2011;124:381-7.

  8. Liu Y et al. J Am Coll Surgeons. 2016;223(2):231-9.

  9. Cohn SL, Fernandez Ros N. Am J Cardiol. 2018;121(1):125-30.

  10. Wijeysundera D et al. Lancet. 2018;391(10140):2631-40.